I've seen this on Windows but not unix. When I wrote the mass test
script for openejb releases I had to sleep for 30s after the test
process exited so the OS would reclaim resources. My guess is if you
simply retry the failure after sleeping for 10s, the test would
pass. IIRC this is the first real MDB test in the suite, and I
suspect that the 10s setup delay (after failure) will give the OS
enough time to catch up.
-dain
On Oct 12, 2007, at 12:09 AM, Jacek Laskowski wrote:
On 10/12/07, Manu George <[EMAIL PROTECTED]> wrote:
I too have noticed this happening rarely. Mostly its not repeatable.
Since it deals with messaging and it may be null because a message
timed out maybe increasing the timeout in MdbProxy will help.
Thanks Manu. I'm looking into it.
Jacek
--
Jacek Laskowski
http://www.JacekLaskowski.pl
